The Catecholamines up (Catsup) Protein of Drosophila melanogaster Functions as a Negative Regulator of Tyrosine Hydroxylase Activity

Corresponding author: Dean G. Stathakis, University of California, Department of Developmental and Cell Biology, 4150 Biological Sciences II, Irvine, CA 92697-2275., dgstatha{at}uci.edu (E-mail)

Communicating editor: V. G. FINNERTY

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We report the genetic, phenotypic, and biochemical analyses of Catecholamines up (Catsup), a gene that encodes a negative regulator of tyrosine hydroxylase (TH) activity. Mutations within this locus are semidominant lethals of variable penetrance that result in three broad, overlapping effective lethal phases (ELPs), indicating that the Catsup gene product is essential throughout development. Mutants from each ELP exhibit either cuticle defects or catecholamine-related abnormalities, such as melanotic salivary glands or pseudotumors. Additionally, Catsup mutants have significantly elevated TH activity that may arise from a post-translational modification of the enzyme. The hyperactivation of TH in Catsup mutants results in abnormally high levels of catecholamines, which can account for the lethality, visible phenotypes, and female sterility observed in these mutants. We propose that Catsup is a component of a novel system that downregulates TH activity, making Catsup the fourth locus found within the Dopa decarboxylase (Ddc) gene cluster that functions in catecholamine metabolism.

THE biosynthesis of all catecholamines initiates with the hydroxylation of L-tyrosine into 3-4-dihydroxy-L-phenylalanine (DOPA), a reaction catalyzed by tyrosine 3-hydroxylase (TH; reviewed in SMEETS and REINER 1994 ). This first step is also the rate-limiting step and is considered the pivotal control point in the catecholamine biosynthetic pathway (reviewed in KUMER and VRANA 1996 ). This is in accordance with the complex array of regulatory mechanisms that direct both the synthesis of TH and its subsequent catalytic activity (reviewed in KUMER and VRANA 1996 ). The level and type of TH protein are controlled by transcriptional regulation, alternative splicing, transcript stability, and translational regulation. After synthesis, the enzymatic activity of TH is modulated by two forms of feedback inhibition, allosteric regulation, enzyme stability, and protein phosphorylation.

Much of the framework for the regulation of TH activity and catecholamine biosynthesis has been elaborated through biochemical, molecular, and pharmacological studies in mammalian cell culture systems (reviewed in KUMER and VRANA 1996 ). However, the genetic and molecular techniques available in the Drosophila model system offer an additional avenue of approach to elucidate and identify proteins that regulate TH activity and catecholamine biosynthesis (reviewed in WRIGHT 1987 ; RESTIFO and WHITE 1990 ). Encoded by the pale (ple) locus (NECKAMEYER and WHITE 1993 ), Drosophila TH shares ~50% amino acid identity to its mammalian orthologs (NECKAMEYER and QUINN 1989 ). In addition, Drosophila TH is regulated by transcriptional and post-translational mechanisms that appear to be similar to those in mammals (NECKAMEYER and QUINN 1989 ; BIRMAN et al. 1994 ), suggesting that the mechanisms directing this regulation may also be retained. These close similarities suggest that a genetic analysis of catecholamine regulation in Drosophila should be a fruitful avenue for uncovering further details of catecholamine regulatory mechanisms and functions that will have general applicability.

In addition to TH, >45 genes have been identified in Drosophila that appear to play a role in catecholamine metabolism (reviewed in WRIGHT 1987 ). Of particular interest is a densely populated cluster of at least 21 genes surrounding the Dopa decarboxylase (Ddc) locus, the ortholog of mammalian AADC. The Ddc gene cluster seems to represent a large grouping of functionally related genes, because mutations in most loci result in similar morphogenetic defects of the cuticle or catecholamine-related abnormalities (STATHAKIS et al. 1995 ; WRIGHT 1996 ). Seven of these genes are of particular interest because HPLC studies revealed that mutations in any of these loci result in mutants with aberrant catecholamine pool levels during prepupal and pupal development (T. HOMYK, W. E. MCIVOR and T. R. F. WRIGHT, unpublished data; cited in WRIGHT 1996 ). One of these genes is Ddc itself (WRIGHT et al. 1982 ), while a second is diphenol oxidase A2 (DOX-A2), a protein that may regulate diphenol oxidase activity, another enzyme involved in catecholamine biosynthesis (PENTZ et al. 1986 ; PENTZ and WRIGHT 1986 ; KAWAMURA et al. 1996 ). A third locus, a-methyl DOPA (amd), encodes a protein structurally related to DDC (EVELETH and MARSH 1986 ; BLACK et al. 1987 ) which may also function as a decarboxylase in catecholamine metabolism (KONRAD et al. 1993 ; WANG and MARSH 1995 ). We report here the results of the genetic, phenotypic, and biochemical analyses of a fourth locus in this group, Catecholamines up (Catsup), so named because one of its mutant phenotypes is an elevation in catecholamine pools. We demonstrate that Catsup, the protein encoded by this gene, acts as a negative regulator of TH activity and thereby defines a new mode of regulation for catecholamine production.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains:
The locus l(2)37Bc (STATHAKIS et al. 1995 ) has been renamed Catecholamines up (Catsup). The P-element insertion strain P{lacW}Catsup^k05402 was kindly provided by Paul Lasko. The OvoD1 (B2121) and T(2;3)ap^Xa (B3234) strains were obtained from the Bloomington Stock Center, and the odd (1699) strain was obtained from the Mid-America Drosophila Stock Center. All other Drosophila melanogaster mutations, aberrations, and balancer chromosomes used were described previously (LINDSLEY and ZIMM 1992 ; STATHAKIS et al. 1995 ).

Genetic testing for lethality:
All crosses mated 12 males to 8 virgin females, and in most cases reciprocal crosses were set up. Hemizygous viability was tested over Df(2L)TW130 and Df(2L)OD15, both of which delete the Catsup locus (STATHAKIS et al. 1995 ). Crosses were scored for the presence of non-Curly O progeny on days 14, 16, and 18. The severity of lethality was classified as follows: v, viable, 86% or more of expected progeny survive to eclose as adults; sv, subvital, 51–85% of expected progeny survive to eclose as adults; sl, semilethal, 6–50% of expected progeny survive to eclose as adults; l, lethal, 5% or less of expected progeny survive to eclose as adults.

Genetic testing for female sterility:
Heteroallelic females were generated by mating 12 Catsup^x/CyO males to 8 Catsup^y/CyO virgin females and collecting the resulting Catsup^x/Catsup^y virgin females. Hemizygotes were generated in a similar fashion, except that Catsup^x/CyO males were mated to female virgins carrying either Df(2L)TW130 or Df(2L)OD15. Female sterility was tested by mating 12 Oregon-R males to 10 Catsup heteroallelic heterozygous virgin females and scoring pupal stage progeny on day 7. To account for possible developmental delays, crosses were monitored for an additional 10 days. Fertility was classified as follows: ++, fully fertile (21 pupae), 0–5% reduction in female fertility; +, fertile (16–20 pupae), 6–25% reduction in female fertility; rf, reduced fertility (11–15 pupae), 26–50% reduction in female fertility; ss, semisterile (6–10 pupae), 51–75% reduction in female fertility; vs, virtually sterile (2–5 pupae), 76–95% reduction in female fertility; fs, female sterile (0–1 pupa), 96–100% reduction in female fertility.

P-element plasmid rescue:
Genomic DNA was purified from P{lacW}Catsup^k05402 adult flies using Qiagen tips, and ~5.0 µg of genomic DNA was digested with SstII (Life Technologies, Grand Island, NY). Digested DNA (~1.5 µg) was added to a 200-µl ligation reaction mix containing 2 units of T4 DNA ligase (Life Technologies), and the reaction was incubated at 14° overnight. Ligated DNA was precipitated in 0.3 M sodium acetate and 2 volumes absolute ethanol, microfuged for 15 min and resuspended in TE buffer. Resuspended DNA was transformed into DH5 max efficiency competent cells (Life Technologies), and plasmid-rescued colonies were selected on 100 µg/ml ampicillin Luria broth (LB) agar plates. Alkaline-lysis-purified plasmid DNA isolated from individual colonies was digested with SstII, and P{lacW} transposon-positive clones were sequenced using the Thermo Sequenase Radiolabeled Terminator cycle sequencing kit (Amersham Life Sciences, Arlington Heights, IL).

P-element mutagenesis screen for imprecise excisions:
The PlacW transposon was remobilized by crossing P{lacW}Catsup^k05402 virgin females en masse to males carrying the P transposase (Figure 1). Individual F₁ male progeny exhibiting somatic mosaicism in the compound eyes were isolated and mated to four virgin females from the T(2;3)ap^Xa balancer stock. Individual F₂ male progeny having lost the w⁺ marker of the P{lacW} transposon were backcrossed individually to eight Catsup¹² virgin females and scored for lethality. Crosses that did not produce viable heteroallelic heterozygous non-Curly O progeny in the next generation were considered potential Catsup mutations. A stock was established for all potential Catsup mutations, and each line was subjected to restriction fragment length polymorphism (RFLP) analysis to determine the extent of the imprecise excision event.

View larger version (11K): [in this window] [in a new window] [Download PPT slide]   Figure 1. Genetic screen to isolate imprecise excisions of P{lacW} transposon in P{lacW} Catsupk05402. The genotype of P{lacW}Catsupk05402 is y w67c23(2)/y w67c23(2), Catsupk05402 P{lacZP/T.W w+mC ampR ori}/CyO. Genotypes of all other fly strains used in this screen are indicated in the figure.

RFLP analysis:
Genomic DNA from 30 adult flies was purified by a potassium acetate/ethanol precipitation procedure. Purified genomic DNA was digested with PstI, electrophoresed, and transferred to a Zeta-probe nylon membrane (Bio-Rad, Richmond, CA) using an alkaline transfer method (REED and MANN 1985 ) modified as follows: an acid pretreatment step was not used, and 0.4 N NaOH served as the transfer solution. Catsup gene-specific ³²P-labeled probes were synthesized from the 1.74-kb EcoRI-HindIII fragment from cDNA Bc541 (STATHAKIS et al. 1995 ) using the Prime-It II random primer labeling kit (Stratagene, La Jolla, CA). High stringency hybridizations and washes were carried out at 65° using the high SDS method (CHURCH and GILBERT 1984 ). Membranes were exposed to Kodak XAR-5 film at -70° with intensifying screens for up to 24 hr before development.

Effective lethal phase analysis:
The effective lethal phase of Catsup alleles was quantitatively measured by the direct method (SHEARN 1978 ), using an egg-picking procedure (SPARROW and WRIGHT 1974 ) with the following modifications. Cultures were established on egg-laying plates and transferred to fresh plates every 2 hr. Eggs were collected and transferred to new egg-laying plates in groups of ~50 and scored every day for viability until eclosion. Mortality at each stage was expressed as a percentage of the total number of fertile eggs. Embryos, larvae, pupa, and adults were also examined for any mutant phenotype. or^49h, a recessive visible mutant allele, was used to mark the Catsup mutant chromosome and phenotypically distinguish homozygous from heterozygous animals (ASHBURNER 1989 ).

Germline clone analysis:
Germline clones of Catsup¹ were generated using the dominant female sterile technique (PERRIMON and GANS 1983 ; PERRIMON et al. 1984 ). Balanced Catsup¹ females were crossed en masse to males carrying two P-element insertions of the dominant female sterile mutation OvoD1 {P[OvoD1](28A, 30D)}. Resulting Catsup¹/OvoD1 virgin females were irradiated with 1200 rad of gamma radiation from a ¹³⁷Cs source. Irradiated females were mated to Catsup¹/CyO males en masse and placed on egg-laying plates. Plates were changed every 24 hr and scored for the presence of eggs.

Light and electron microscopy analysis:
Dissected pupae and pharate adults were divided into head, thoracic, and abdominal regions and fixed for 24 hr in 100 mM sodium phosphate (pH 7.2) containing 2.5% glutaraldehyde and 4% paraformaldehyde. Samples were washed twice and postfixed in 100 mM sodium phosphate containing 2% osmium tetroxide for 3 hr, dehydrated in acetone, and embedded in Epon 812. For light microscopy, both longitudinal and cross semithick sections (0.5 µm) were cut and stained with either 0.5% toluidine blue (5 min) or 2% acid fuchsin (10 min). For transmission electron microscopy, samples were trimmed and thin gold sections were collected on Formvar-coated copper slot grids, poststained with 5% uranyl acetate and 0.25% lead citrate, and examined in a JEOL 100CX electron microscope.

Phalloidin staining with confocal microscopy analysis:
Adult females were dissected in Ringer's solution, and their ovarioles were fixed in Ringer's solution containing 4% formaldehyde for 60 min. Ovarioles were washed three times for 5 min each time in 1x PBS (pH 7.4) and incubated in blocking buffer [1x PBS (pH 7.4), 2% BSA, 0.2% NP-40, 0.02% sodium azide] for 30 min containing a 1:100 dilution of fluorescein-conjugated phalloidin (Molecular Probes, Eugene, OR) to visualize filamentous actin. Tissues were washed four times for 15 min each time in 1x PBS (pH 7.4) and mounted in Vectashield (Vector Laboratories, Burlingame, CA) to limit photobleaching. Tissues were visualized using an MRC 640 confocal microscope (Bio-Rad).

High performance liquid chromatography analysis:
Pool sizes of tyrosine and catecholamines were determined during late embryogenesis by reverse-phase high performance liquid chromatography (RP-HLPC) with electrochemical detection. Staged dechorionated embryos were sorted by phenotype using the odd-skipped allele odd^IIID as an embryonic marker (NUSSLEIN-VOLHARD et al. 1984 ). Sorted embryos were immediately homogenized in 100 mM perchloric acid, 4 mM EDTA (50 embryos in 100 µl buffer), chilled on ice for 15 min to precipitate proteins, and centrifuged for 10 min at 27,000 x g at 4° to remove tissue debris. Samples were immediately stored at -70° until needed.

Chromatographic separations were carried out on either a Beckman 334 liquid chromatography apparatus equipped with a 100A pump and 427 integrator or a Beckman 126 liquid chromatography apparatus equipped with the System Gold software package (Beckman, Fullerton, CA). Samples were loaded using a Spectrophysics SP8780 autoinjector fitted with a 10-µl injection loop. Separations were achieved using a 4.6 x 100-mm, 3-µm reverse-phase ODS particle column (Ranin Microsorb Short One) using an ESA Coulochem II EC detector fitted with a 5021 conditioning cell and a 5011 analytical cell. The conditioning cell was set at -0.04 V, while the two detector electrodes were set at +0.25 V to detect catechols and hydroxyindoles and +0.55 V to detect hydroxyphenyl compounds and indoles with substituted rings. Column temperature was held constant at 30°. The most commonly used mobile phase consisted of 75 mM sodium phosphate (pH 3.0), 1.7 mM octanesulfonic acid, 0.01% (v/v) triethylamine, 0.025 mM EDTA, and 11% (v/v) acetonitrile at a flow rate of 1.00 ml/min. Column retention times and linear detector response were calibrated using commercial standards, except for N-ß-alanyl dopamine (NBAD), N-acetyl norepinephrine (NANE), and N-ß-alanyl norepinephrine (NBANE), which were kindly provided by Karl J. Krammer. The limit of detection ranged from 10 fmol DOPA to 50 fmol tryptophan per injected sample. Typical standard elution times are as follows:

• +0.25 V detector: DOPA 1.57 min, NANE 1.81 min, NBANE 2.13 min, N-acetyl dopamine (NADA) 2.73 min, normetanephrine 2.98 min, dopamine (DA) 3.65 min, 5-hydroxy-indole acetic acid (5-HIAA) 4.2 min, acetyl serotonin 4.79 min, homovanillic acid (HVA) 6.11 min, and NBAD 7.90 min;

• +0.55 V detector: L-tyrosine 2.00 min, M-tyrosine 2.44 min, normetanephrine 3.08 min, vanillic acid 5.04 min, and tryptophan 13.6 min.

Chromatograms of free catecholamine pool levels routinely resolved five major electroactive compounds in amounts of 1 pmol/embryo, of which three were identified as DOPA, DA, and NADA. Chromatograms of free tyrosine pool levels routinely resolved three electroactive compounds in amounts of 1 pmol/embryo, one of which was L-tyrosine.

Dopamine feeding experiments:
Two different feeding experiments were conducted on media augmented with DA. In the first approach, 70–100 Oregon-R first instar larvae were harvested from 24-hr egg-laying plates and transferred to Petri dishes with filter paper containing a 25% yeast paste supplemented with either 1, 2, 3, 10, or 20 mg/ml DA. Cultures were monitored daily until eclosion, and animals were scored for delays in developmental rates and any visible larval or pupal phenotypes. In the second experimental approach, five Oregon-R virgin females were mated to five Oregon-R males and transferred to instant Drosophila medium (Carolina Biological) reconstituted with water supplemented with either 1, 2, 5, or 10 mg/ml DA. Adult flies were removed after 3 days, and the vials were subsequently monitored for the numbers of progeny that pupated and eclosed and for effects on developmental rates. Adult flies resulting from these crosses were transferred to vials containing unsupplemented media and were scored daily for fecundity and progeny viability.

Tyrosine hydroxylase assays:
The L-tyrosine-3, 5-³H microradioassay method (NAGATSU et al. 1964 ) was used with the following changes. Samples were homogenized in ice-cold 1x PBS (pH 7.0) containing 1 mM phenylmethylsulfonyl fluoride and 1 mM dithiothritol (25 larvae in 300-µl buffer). Debris was removed by centrifugation for 10 min at 4000 x g, and protein concentrations were determined. Two hundred microliters of fly extract, adjusted to 1 µg/µl protein, were added to a 300-µl reaction solution consisting of 200 µl 0.15 mM sodium phosphate (pH 6.7) containing 1950 units of catalase, 50 µl 1.0 mM ascorbic acid, 1.0 mM 6, 7-dimethyl 5, 6, 7, 8-tetrahydropterin (DMPH₄) in 1 mM ascorbic acid, and 50 µl 1.0 mM L-tyrosine with a ³H-tyrosine tracer (100,000 cpm). The reaction components were incubated with shaking at 32° in uncapped, round-bottomed tubes for 25 min. The remainder of the procedure was essentially as described by NECKAMEYER and QUINN 1989 . The enzyme reaction demonstrated linear reaction kinetics with respect to increasing enzyme concentrations.

Western blot analysis:
Total protein was extracted from third instar larvae, separated by 10% SDS-PAGE, and transferred to PVDF filters (Bio-Rad) using standard methods (HARLOW and LANE 1988 ). Filters were blocked for 5 hr in 50 mM Tris-HCl (pH 7.6), 100 mM NaCl, 0.2% Tween-20, 5% nonfat dry milk, and incubated overnight in a 1:2500 dilution of anti-TH mouse monoclonal antibody (Pel-Freez), a 1:17,000 dilution of guinea pig polyclonal anti-Catsup antiserum, or a 1:15,000 dilution of anti--tubulin mouse monoclonal antibody (Sigma, St. Louis, MO). Blots were washed three times for 30 min each time in 50 mM Tris-HCl (pH 7.6), 100 mM NaCl, 0.2% Tween-20, incubated for 1 hr in a 1:2000 dilution of anti-mouse or a 1:2000 anti-guinea pig IgG-HRP-conjugated secondary antiserum (Jackson Immunological), and washed three times for 30 min in 50 mM Tris-HCl (pH 7.6), 100 mM NaCl, 0.2% Tween-20. Signals were visualized by the Renaissance chemiluminescence reagent (Dupont NEN) and demonstrated a linear response under the conditions used. Western blot films were scanned using a BioImage analyzer and Visage software. The signal in each band was quantitated as the integrated optical density of the band (OD x mm²). Relative band densities were normalized to protein loads as determined by the band density of the control -tubulin in each lane.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Genetic analysis of Catsup
Characterization of Catsup mutations: Twenty-nine mutant alleles of Catsup were recovered as part of a saturation mutagenesis for lethal and sterile mutations within the Ddc gene cluster (STATHAKIS et al. 1995 ). As hemizygotes, 25 Catsup alleles are nonconditional lethals and designated Catsup¹ to Catsup²⁵. Catsup^ts1 is a temperature-sensitive lethal with a restrictive temperature of 29° (STATHAKIS 1998 ). At permissive temperatures (25° or lower), homozygous Catsup^ts1 females are fully fertile. Catsup^cs1 and Catsup^cs2 are cold-sensitive lethals with restrictive temperatures of 22° or lower and permissive temperatures of 25° or higher (STATHAKIS 1998 ). Both alleles exhibit extremely reduced viability at 22° and complete lethality at 18°. Hemizygous escaper females for either allele are completely sterile. The last allele, Catsup^fs1, is a female sterile. Catsup^fs1 hemizygotes are completely viable but female sterile at all experimental temperatures tested (18°, 22°, 25°, and 29°).

Catsup mutations can be ordered into an allelic series: The 25 nonconditional lethal Catsup mutations can be divided into five groups on the basis of the penetrance of lethality and female sterility (Table 1). Group I (GI) and group II (GII) lethal mutations have marginal effects on viability and fertility and are considered weak Catsup alleles. A single lethal allele, Catsup¹¹, is considered a GI lethal because it shows reduced viability as a hemizygote, behaving as a semilethal mutation with hemizygous female escapers being completely sterile (Table 2). This contrasts with GII lethals (Catsup¹⁴, Catsup¹⁷, and Catsup¹⁹) that are hemizygous lethal. Homozygous mutants of either GI or GII lethal alleles exhibit incomplete penetrance for lethality, but differ in their effects on female fertility (Table 2). Catsup¹¹ homozygous adult female escapers are fertile, while Catsup¹⁴, Catsup¹⁷, and Catsup¹⁹ homozygous female escapers exhibit drastically reduced fertility (Table 2). In addition, while GI/GII heteroallelic heterozygotes are completely fertile, GII/GII heteroallelic heterozygotes are female sterile (Table 2).

Group III (GIII) and group IV (GIV) lethal mutations have moderate effects on progeny viability and fertility. The 14 Catsup alleles comprising these two categories are lethal as hemi- and homozygotes (Table 2). However, these alleles can be separated into two categories on the basis of their interactions with GII lethal mutations and as heteroallelic heterozygotes with respect to each other. The seven alleles comprising GIII lethal mutations (Catsup², Catsup⁶, Catsup⁸, Catsup⁹, Catsup¹⁵, Catsup¹⁸, and Catsup²³) produce progeny when heterozygous with any GII lethal allele (Table 2). This contrasts with GIV lethal mutations (Catsup⁴, Catsup⁵, Catsup⁷, Catsup¹⁰, Catsup¹², and Catsup²⁴), which are lethal when heterozygous with any GII lethal allele (Table 2). In addition, except for Catsup¹⁵, GIII/GIII heteroallelic heterozygotes are viable and fertile, while all GIV/GIV heteroallelic heterozygotes are lethal (Table 2).

Allele-specific interactions are evident for GIII lethal mutations. Among GIII lethal mutations, Catsup¹⁵/Catsup⁶, Catsup¹⁵/Catsup⁸, and Catsup¹⁵/Catsup⁹ heterozygotes are viable, while Catsup¹⁵/Catsup², Catsup¹⁵/Catsup¹⁸, and Catsup¹⁵/Catsup²³ heterozygotes are lethal. This indicates that Catsup¹⁵ is a strong negative complementer of the latter three alleles. Likewise, Catsup⁸ negatively complements Catsup² and Catsup⁶, as heterozygotes have dramatically reduced viability. This negative complementation suggests that GIII lethal mutations might affect two different functional aspects of the Catsup gene product. Allele-specific interactions also occur between GIII/GIV heterozygous combinations, suggesting that GIV lethal mutations are not complete loss-of-function mutations. These interactions are most evident with the group III lethal mutation Catsup⁹. While other GIII lethals fail to complement GIV lethals, Catsup⁹ complements all GIV lethal mutations, except for Catsup¹² (Table 2). The majority of GIII heteroallelic heterozygotes are female sterile (Table 2). However, heterozygotes between Catsup⁹ and most GIII lethal mutations are fully fertile. Only Catsup⁹/Catsup¹⁵ heterozygotes are completely sterile (Table 2). GIV lethal mutations have a slightly more severe effect on female fertility because all GIV/GIV heteroallelic heterozygotes are completely sterile (Table 2). This contention is further supported by interallelic interactions of GIII and GIV lethal alleles with the female-sterile allele Catsup^fs1. Catsup^fs1/GIII heteroallelic heterozygotes exhibit a wide phenotypic range, with Catsup^fs1/Catsup¹⁵ and Catsup^fs1/Catsup²³ female adults displaying full fertility (Table 2). In contrast, all Catsup^fs1/GIV heteroallelic heterozygotes show severely reduced fertility, with the vast majority of these heterozygotes being completely female sterile (Table 2).

The seven remaining Catsup mutations (Catsup¹, Catsup³, Catsup¹³, Catsup¹⁶, Catsup²¹, Catsup²², and Catsup²⁵) are considered group V (GV) lethal mutations as they have the most severe effects on viability and fertility. Heterozygotes between a GV lethal allele and any other Catsup allele, except for the two weakest mutations, Catsup^fs1 and Catsup¹¹, are lethal (Table 2). Heterozygotes for Catsup¹¹ or Catsup^fs1 and a GV lethal allele are always female sterile. The results with GV lethal alleles are identical to those obtained from Df(2L)TW130 and Df(2L)OD15, two different deficiencies that delete the Catsup locus (Table 2). This suggests that GV lethals are null mutations.

Taken together, this analysis indicates that the nonconditional lethal alleles of Catsup can be grouped in terms of increasing penetrance of lethality and female sterility. GI and GII lethals have weak effects, and GIII and GIV lethals display moderate effects. GV lethals have the most severe effects, which behave similarly to deletions of Catsup.

Generation of Catsup deletion mutation: The Catsup locus lies within the distal subcluster of the Ddc gene cluster (STATHAKIS et al. 1995 ). Both Df(2L)TW130 and Df(2L)OD15 remove many loci within the Ddc gene cluster, several of which also affect catecholamine pool levels. Therefore, there was concern that aspects of the Catsup null phenotype may have been altered or masked by the presence of only one copy of one or several of these other loci involved in catecholamine metabolism. To address this issue, we conducted a P-element mutagenesis screen to generate imprecise excision mutations to generate a deletion within the Catsup locus.

We obtained a Berkley Drosophila Genome Project P-element insertion strain (SPRADLING et al. 1995 ), designated P{lacW}Catsup^k05402. P{lacW}Catsup^k05402 behaves genetically as a severe GV lethal allele of Catsup. P{lacW}Catsup^k05402 mutations are lethal when heterozygous with the moderate GIV lethal allele Catsup¹², the severe GV lethal allele Catsup¹, or Df(2L)TW130 and Df(2L)OD15. In addition, the penetrance of the weak GI lethal mutation Catsup¹¹ is increased to similar levels when in trans to either P{lacW}Catsup^k05402, the severe GV lethal allele Catsup¹, or Df(2L)TW130 and Df(2L)OD15. We cloned the genomic DNA region flanking the insertion site of the P{lacW} transposon in P{lacW}Catsup^k05402 by plasmid rescue. Subsequent sequence analysis indicated that the P{lacW} transposon is inserted in the 5' UTR of the Catsup transcription unit, 26 bp upstream from the AUG initiator methionine codon (Figure 2).

View larger version (9K): [in this window] [in a new window] [Download PPT slide]   Figure 2. RFLP analysis of Catsup imprecise excision mutations. The figure depicts the restriction endonuclease map of the genome region surrounding the Catsup locus. The genomic organization of the Catsup locus consists of two exons (STATHAKIS 1998 ) and is illustrated as follows: cross-hatched rectangles represent the protein coding regions of the exons, hatched rectangles represent the 5' and 3' UTRs of the exons, and the white rectangle represents the single intron. Also shown is the insertion site of the P{lacW} transposon within the 5' UTR of Catsup. The location of the transcription units for l(2)37Be, Catsup, and l(2)37Ba are represented by black rectangles. The two PstI DNA fragments recognized by the Catsup cDNA probe in the RFLP analysis are indicated below the restriction map. Restriction enzyme symbols are as follows: Bg, BglII; EI, EcoRI; Hd, HindIII; Ps, PstI; Ss, SstII. The diagram is drawn to scale.

To generate deletion mutations in the Catsup locus, we screened for imprecise excisions of the P{lacW} transposon in P{lacW}Catsup^k05402 by using the 2-3 transposase (Figure 1). Individual strains from 56 independent imprecise excision chromosomes were isolated, and males were backcrossed to Catsup¹² females. Flies from 21 of these isolated lines proved lethal in trans to Catsup¹². These lines were considered new Catsup mutations and designated Catsup²⁶–Catsup⁴⁶ (Table 3). Flies from each new allele were screened for disruption of the Catsup locus by RFLP analysis to determine the extent of the imprecise excision. A total of 16 lines exhibited visible polymorphisms relative to the 13.7-kb parental fragment that contains the 10.6-kb P{lacW} transposon of strain P{lacW}Catsup^k05402, while 5 lines showed no visible size reduction of the 13.7-kb parental fragment (Table 3). Of the 16 lines that did possess a visible polymorphism, the extent of the imprecise excision in 14 of these isolates ranges from 2.6 to 10.3 kb (Table 3). Because the P{lacW} transposon is 10.6 kb, these data suggest that these 14 lines still contain a remnant of the 10.6-kb transposon. Thus, we consider these 14 lines, in conjunction with the 5 isolates that show no visible polymorphism, as insertion mutations of Catsup (Table 3). The remaining two isolates (Catsup²⁶ and Catsup²⁷) show a clear reduction in the 13.7-kb band to a size that is less than the 3.1-kb wild-type genomic DNA fragment. This suggests that in these two lines, regions of the Catsup locus in addition to the entire 10.6-kb P{lacW} transposon were removed. The first isolate, Catsup²⁶, exhibits an ~600-bp deletion of genomic DNA, while the second isolate, Catsup²⁷, shows a smaller deletion of ~300 bp. Because the excision event in both isolates does not affect the function of either locus flanking Catsup (see below), we deduce that Catsup²⁶ and Catsup²⁷ are deletions in the Catsup locus.

Two distal subcluster genes, l(2)37Be and l(2)37Ba, are located <500 bp on either side of the Catsup locus (STATHAKIS et al. 1995 ; see Figure 2). The close proximity of l(2)37Be and l(2)37Ba relative to Catsup raised concerns that an imprecise excision event of the P{lacW} transposon could also have removed essential regions of these two flanking transcription units. To determine whether potential Catsup mutations affected the function of either flanking gene, we conducted crosses with l(2)37Be⁴ and l(2)37Ba¹. Of the 22 Catsup mutations generated by this screen, three isolates (Catsup⁴⁴, Catsup⁴⁵, and Catsup⁴⁶) were also lethal for l(2)37Be⁴ (Table 3). None of the 22 Catsup mutations were lethal for l(2)37Ba¹.

Catsup²⁶ was selected for further analysis because this mutation represents the largest deletion of the Catsup locus that did not affect the function of l(2)37Be and l(2)37Ba. With respect to genetic interactions, ELPs, fertility, and visible phenotypes, Catsup²⁶ is indistinguishable from other GV lethal alleles. In addition, in every aspect examined, the lethal, sterile, and visible phenotypes of Catsup mutations were identical when tested with Catsup²⁶, Df(2L)TW130, and Df(2L)OD15. Catsup²⁶ mutations are lethal when heterozygous with Catsup¹² (GIV) or Catsup¹ (GV) lethal alleles or deficiencies Df(2L)TW130 and Df(2L)OD15. In addition, the penetrance of the weak lethal mutation Catsup¹¹ (GI) is increased to similar levels when in trans to either Catsup²⁶, the GV lethal allele Catsup¹, or the deficiencies Df(2L)TW130 and Df(2L)OD15. This analysis clearly indicates that GV lethal alleles are genetic nulls, and that all the observed lethal, sterile, and visible phenotypes can be solely attributed to mutations in the Catsup locus.

Catsup nonconditional mutations are semidominant lethals that exhibit three broad ELPs: To elucidate when the Catsup gene product is required during development, we determined the ELP for each of the five nonconditional lethal categories of Catsup mutants. In conjunction with our phenotypic examination, this analysis indicates that the onset of lethality can be divided into three broad periods and that a distinctive phenotype is associated with each ELP (Table 4).

Early larval lethals: The first ELP occurs between the late embryonic period and the second larval instar. Approximately 65–75% of the total lethality observed in Catsup mutants hemi- or homozygous for the moderate lethal alleles Catsup²³ (GIII) and Catsup¹² (GIV) and the severe lethal allele Catsup¹ (GV) takes place in this ELP (Table 5). However, while the overall level of total lethality is comparable among these three lethal groups, the relative distribution of lethal onset between each of the three developmental periods is different. The lethality seen in larvae hemi- or homozygous for the Catsup¹ (GV) allele is more evenly distributed, with total observed lethality occurring in ~22% of the embryos, 30% of the first instars, and 23% of the second instars (Table 5). In larvae hemi- or homozygous for Catsup¹² (GIV), lethality occurs primarily in the first instar, with total observed lethality occurring in ~20% of the embryos, 44% of the first instars, and 10% of the second instars (Table 4). The distribution of lethality arises even later in hemi- or homozygous Catsup²³ (GIII) mutants, with total lethality occurring in ~8% of embryos, 28% of first instars, and 30% of second instars (Table 5).

The most common phenotype associated with early larval lethality is arrested development. Hatched Catsup mutant first instar larvae may survive up to 4 days, but do not grow or molt. In addition, these stunted larvae are akinetic, crawling slowly across the food or moving only when prodded, and may also exhibit patches of internal melanization. Likewise, the majority of dying embryos undergo internal melanization several hours before hatching. However, this late embryonic lethality cannot be attributed solely to their inability to hatch because the majority of manually dissected, living larvae are akinetic. Larvae that reach second instar are of normal size, but they exhibit the same akinetic behavior and internal patches of melanization seen in Catsup mutant first instar larvae. These second instar larvae may survive up to 2 days, but do not grow or molt.

Late larval lethals: The second ELP occurs during the third instar, although some lethality is seen earlier in second instars. Approximately 35–40% of the total lethality seen in hemi-, homo- and heterozygotes for the weak lethal alleles Catsup¹¹ (GI) and Catsup¹⁹ (GII) occurs in this ELP. In addition, this ELP accounts for ~12–17% of the total lethality seen in Catsup mutants heterozygous for the moderate lethal alleles Catsup²³ (GIII) and Catsup¹² (GIV), and the severe lethal allele Catsup¹ (GV; Table 5). The death of ~10–20% of either balanced or nonbalanced third instar heterozygotes indicates that Catsup mutations are semidominant lethals (Table 5). Surprisingly, this lethality occurs in heterozygotes from any of the five lethal groups.

Regardless of the allele, dying third instar larvae manifest a very distinctive "melanotic salivary gland" phenotype (Figure 3). In these larvae, individual cells of the salivary glands begin to melanize until, in the most common form, the entire gland is a completely black, mulberry-shaped organ. In the more severe examples, a secondary site of melanization occurs in the mid-dorsal region, probably in the lymph glands, which then spreads throughout the hemolymph. This phenotype can be expressed at any time during third instar.

View larger version (112K): [in this window] [in a new window] [Download PPT slide]   Figure 3. Late larval lethal phenotypes associated with Catsup mutations. (A) Onset of melanotic salivary gland phenotype. Homozygous Catsup11 larvae from left to right: second instar, early third instar, mid-third instar, late third instar and white prepupa. There is abnormal melanization centered around the lymph area (arrows) in the posterior third of the last two larvae. (B and C) Phase-contrast light microscopy of dissected salivary glands from homozygous Catsup11 third instar larvae.

Pupal lethals: The third ELP exhibits the most variable time of onset, occurring at any time during pupariation (Table 4). Total lethality occurs in ~45–55% of pupae hemi-, homo-, and heterozygotes for the weak lethal alleles Catsup¹¹ (GI) or Catsup¹⁹ (GII), and ~15–20% of pupae heterozygotes for the moderate lethal alleles Catsup²³ (GIII) and Catsup¹² (GIV) and the severe lethal allele Catsup¹ (GV, Table 5). As observed in the second ELP, lethality occurs in ~10–20% of heterozygotes from any of the five lethal groups.

The Catsup mutant phenotype during prepupal development is characterized by large regions of internal melanization in prepupae and pupal cases that are darker than normal, especially after aging (Figure 4A). In addition, most prepupae that die start autolysis but cannot undergo metamorphosis, producing empty or partially empty pupal cases. Both Catsup mutant pupae and pharate adults undergo severe, abnormal melanization and display varying degrees of incomplete cuticle formation. The cuticle is relatively normal on the head and thorax, but is incompletely formed over much of the abdomen (Figure 4B). No bristles are formed in the abdomen, and the abdomen shows no indications of either segmentation or normal patterns of tanning. Furthermore, the abdomens of these pharate adults are usually extensively melanized in a blotchy pattern. While the head, thoracic, and leg cuticles appear to be formed normally, the thoracic cuticle in particular is not completely hardened. Initial inspection of the gross morphology by light microscopy reveals that Catsup pharate adults appear to have normal cuticles, but they lack an epidermis in regions exhibiting incomplete cuticles. Further examination by electron microscopy indicates that while thinner relative to control pharate adults, the cuticles of Catsup mutants have both well-defined epicuticles and procuticles with normal appearance (Figure 4E and Figure F).

View larger version (109K): [in this window] [in a new window] [Download PPT slide]   Figure 4. Pupal lethal phenotypes associated with Catsup mutations. (A) Comparison of wild-type (left two) and Catsup11/Df(2L)TW130 mutant (right two) pupae. There are internal regions with intense melanization within Catsup mutant pupae. (B) Dissected Catsup11/Df(2L)TW130 late pharate adult. There is an absence of tergite and bristle formation in the posterior abdominal region. Melanization is apparent through the undamaged pupa case at the time of imaginal disc eversion and darkens as development proceeds. (C and D) Micrographs of pharate adult abdominal cuticle. Light micrographs (x2000) of serial sections stained with toluidine blue in a control (C) and Catsup11 or49h homozygote (D). Controls show well-defined epidermis (EPI) and cuticle (CUT). There is a complete absence of a dark-staining epidermal layer in Catsup11 mutants, but a CUT is present. The lack of internal organs or tissues in these mutants also suggests that the disappearance of the epidermis may result from necrosis. All sections are x2000, oil immersion. (E and F) Transmission electron micrographs (x16,000) of cross-sections of a control (E) and a Catsup11 or49h homozygote (F). Controls show a well-formed, simple cuboidal epithelium (EPI) and the two subdivisions of the cuticle, an electron-dense epicuticle (EC), and a distinctly lamellar procuticle (PC). While Catsup11 mutants exhibit no epidermis, cuticle structure appears normal, albeit thinner relative to controls, because both well-formed epicuticle and procuticle layers are present.

Catsup mutant adult escaper phenotypes: Incomplete cuticle formation, melanotic pseudotumors, and female sterility are associated with Catsup mutant adult escapers. Approximately 30–50% of the adult escapers produced by crosses between balanced GI and GII lethal mutations, as well as various heteroallelic combinations of GIII, GIV, and GV lethal mutations, exhibit minor malformations of the cuticle, primarily on the abdominal tergites (Figure 5, A–D). Another phenotype associated with Catsup adult escapers is melanotic pseudotumors. These acellular tumors occur in ~5–10% of Catsup³, Catsup⁵, Catsup⁶, Catsup⁸, Catsup¹¹, Catsup¹³, Catsup²², Catsup^fs1, Catsup^cs1, and Catsup^cs2 balanced heterozygotes. Although present in only a subset of Catsup alleles, there is no association with mutation severity because these 10 alleles represent weak, moderate, and severe lethal mutations, as well as the female-sterile and both cold-sensitive mutations. In addition, while melanotic pseudotumors may occur in mutants as early as the second larval instar, their formation does not affect viability because larvae will survive to become fertile adults.

View larger version (112K): [in this window] [in a new window] [Download PPT slide]   Figure 5. Adult visible phenotypes associated with Catsup mutants. (A and B) Incomplete formation of abdominal tergites in Catsup11 homozygous female (A) and male (B) adults. (C and D) Incomplete formation of abdominal tergites in Catsup12/Catsupfs1 heterozygous female (C) and male (D) adults. (E and F) Phase-contrast light microscopy of eggs produced by adult females whose germline is homozygous for the Catsup1 mutation.

The most common adult phenotype is female sterility. Female sterility is seen in all adults homozygous for Catsup¹⁴, Catsup¹⁹, or either cold-sensitive allele raised at permissive temperatures. In addition, ~66% of the crosses that produce heteroallelic heterozygotes are female sterile, with an additional 10% exhibiting some loss of fertility (Table 2). Dissected spermatheca of Catsup female-sterile mutants contain large numbers of spermatozoa after mating with wild-type males, indicating that these mutants are inseminated. In addition, female-sterile mutants contain bloated ovarioles containing large numbers of apparently mature, yolk-filled eggs. However, the fecundity of these mutants is dramatically reduced with females depositing only a few small flaccid eggs.

Closer examination of hemizygous Catsup^fs1 females by DIC and confocal microscopy indicates an abnormal retention of eggs, with every ovariole of a 4-day-old adult containing two to five normal-sized oocytes (Figure 6). Approximately 50% of oocytes from these Catsup female-sterile mutants lack dorsal appendages. However, rudimentary filaments do form in the remainder of these oocytes, suggesting that oogenesis can proceed until stage 13. A chorion membrane is always present. The formation of micropyles in addition to the presence of rudimentary dorsal appendages suggests that oocytes from these Catsup female-sterile mutants undergo normal pattern formation. Mature, stage 14 oocytes are never present in ovarioles from these Castup female-sterile mutants. Likewise, ovarioles contain <5% stage 8–11 egg chambers. A similar egg retention phenotype is seen in Catsup^cs1 and Catsup^cs2 hemizygous or Catsup^fs1/Catsup¹ and Catsup^fs1/Catsup¹² Catsup^cs1/Catsup^cs2 heterozygous escaper females.

View larger version (100K): [in this window] [in a new window] [Download PPT slide]   Figure 6. Visible phenotypes associated with oogenesis in Catsup mutants. (A, C, and D) DIC microscopy of ovarioles dissected from a 2- to 4-day-old Oregon-R adult female (A) and from 2- to 4-day-old Catsupfs1 hemizygous adult females (C and D). Although arresting at stage 13 of oogenesis, Catsup mutant ovarioles appear morphologically normal, possessing a micropyle, chorion membrane, and dorsal appendages. (B and E–G) Confocal microscopy of ovarioles stained with fluorescein-conjugated phalloidin to visualize filamentous actin. Ovarioles dissected from 2- to 4-day-old Oregon-R adult female (B) and from 2- to 4-day-old Catsupfs1 hemizygous adult females (E–G). Similar phenotypes are observed in 2- to 4-day-old Catsupcs1 and Catsupcs2 hemizygous adult females and 2- to 4-day-old Catsupfs1/Catsup1 and Catsupfs1/Catsup12 and Catsupcs1/Catsupcs2 heterozygous adult females.

Germline clone analysis of the GV lethal allele Catsup¹ confirms a critical role for the Catsup gene product during oogenesis. Females with homozygous Catsup¹ mutant germlines deposit few, morphologically defective eggs that appear to have been arrested at mid- to late oogenesis (Figure 5E and Figure F). The eggs produced are flaccid when laid or collapse during egg deposition, suggesting the chorion is weak or absent, and the egg is only surrounded by the vitelline membrane. Intact eggs have rudimentary dorsal appendages that are fused to the eggshell, or they lack such appendages all together. Large cuboidal cells, reminiscent of nurse cells, comprise the anteriocentral region of the follicle, with a yolk-filled oocyte occupying the posterior end, a cellular organization similar to that of the "dumpless" phenotype (PERRIMON et al. 1989 ). This indicates that females with homozygous Catsup¹ mutant germlines produce egg chambers with more severe abnormalities relative to hemizygous Catsup^fs1 females.

Biochemical analysis of Catsup
Catsup mutants exhibit elevated catecholamine pool levels during embryogenesis: The aberrant melanization observed in Catsup mutants could be indicative of abnormal regulation of catecholamine pool levels. In Drosophila, the four major catecholamines produced are DOPA, DA, NADA, and NBAD (Figure 7). DOPA serves primarily as a precursor compound in the catecholamine biosynthetic pathway. DA functions as a neurotransmitter (reviewed in RESTIFO and WHITE 1990 ) and is the precursor compound for all subsequent catecholamines used in the melanin and sclerotin biosynthetic pathways (reviewed in HOPKINS and KRAMER 1992 ). NADA and NBAD are the two major cross-linking agents in sclerotization. NADA and its derivatives (dehydro-NADA, NANE) are found primarily in colorless (clear) sclerotin, while NBAD and its derivatives (dehydro-NBAD, NBANE) are present principally in colored (brown) sclerotin (ANDERSON 1989A , ANDERSON 1989B ; HOPKINS and KRAMER 1992 ; SUGUMARAN et al. 1992 ). Therefore, measuring the levels of these four catecholamines should determine whether misregulation of catecholamine biosynthesis is occurring in Catsup mutants.

View larger version (18K): [in this window] [in a new window] [Download PPT slide]   Figure 7. Catecholamine biosynthesis pathway in D. melanogaster. The figure depicts the initial enzymatic steps required to synthesize the four major catecholamines necessary for neurotransmitter production, melanization, and sclerotization (reviewed in WRIGHT 1987 ). The hydroxylation of tyrosine (TYR) to form DOPA is catalyzed by TH. The decarboxylation of DOPA to form DA is catalyzed by DOPA decarboxylase (DDC). The N-acylation of DA with acetate to form NADA is catalyzed by NADA transferase (NAT), while the N-acylation of DA with ß-alanine to form NBAD is catalyzed by NBAD transferase (NBAT). The conversion of all catecholamines to their respective quinones is catalyzed by the phenol oxidase enzyme complex. Loci encoding the enzyme(s) responsible for each reaction are in parentheses, with references for gene assignments as follows: ple, NECKAMEYER and WHITE 1993 ; Ddc, WRIGHT et al. 1976 ; Nat, WRIGHT 1987 ; t and e, BLACK 1988 ; Mox, FUJIMOTO et al. 1995 ; Dox-A2, PENTZ et al. 1986 .

We determined catecholamine pool levels in embryos homozygous for the GIV lethal allele Catsup¹². Because we estimate that ~70% of Catsup¹² homozygous mutants die during the first larval instar, catecholamine pool levels were determined in 22-hr-old embryos, a time point representing peak catecholamine synthesis during embryogenesis (WRIGHT 1987 ; D. G. STATHAKIS, W. E. MCIVOR and T. R. F. WRIGHT, unpublished data). However, because ~30% of Catsup¹² homozygous mutants die during embryogenesis, great care was taken to select embryos that were still alive and did not exhibit patches of internal melanization. Catsup¹² mutant embryos exhibit pronounced increases in DA and NADA pools at 22 hr of embryonic development, indicating that catecholamine biosynthesis is misregulated during embryogenesis (Table 6). Relative to controls, DA pools increase more than fourfold in homozygous Catsup¹² mutants. Similarly, NADA pools increase more than fivefold in Catsup¹² mutant embryos. Interestingly, heterozygous Catsup¹² mutants also exhibit significantly higher levels of DA and NADA relative to controls (Table 6). This increase is approximately midrange between homozygous Catsup¹² mutants and controls, indicating that one normal allele provides a single functional dose of the Catsup gene product.

In contrast to the effects observed for DA and subsequent catecholamines, DOPA pool levels in Catsup¹² mutant embryos are not significantly elevated relative to wild-type levels. Interestingly, both homo- and heterozygous Catsup mutant embryos exhibit significantly reduced levels of free tyrosine relative to controls (Table 6). Tyrosine pools decrease ~3-fold in homozygous Catsup¹² mutants relative to controls. A similar decrease (2.6-fold) of tyrosine pool levels was also observed in heterozygous Catsup¹² mutant embryos (Table 6).

Our findings are consistent with HPLC analysis of Catsup¹¹ (GI) hemizygotes, which showed dramatic increases in catecholamine pool levels at four distinct times during the first 72 hr of pupariation (T. HOMYK, W. E. MCIVOR and T. R. F. WRIGHT, unpublished data, cited in WRIGHT 1996 ). During this period, mutants from this GI lethal allele exhibit ~7.5-fold higher levels of NADA, 4-fold higher levels of NBAD and NBANE, and 3-fold higher levels of DA relative to controls. DOPA levels were unaffected.

Exogenous dopamine phenocopies Catsup mutations: To test directly whether any of the phenotypes associated with Catsup mutations are caused by elevated DA levels, we transferred wild-type first instar larvae to food supplemented with 1, 2, 3, 10, and 20 mg/ml of DA. Many of the phenotypes associated with mutations in Catsup are phenocopied under these conditions and are positively correlated with increasing DA concentrations with respect to severity of the observed phenotypes. At higher DA concentrations, these larvae exhibit several early larval lethal phenotypes associated with Catsup mutants. Larval lethality is evident on media supplemented with 10 or 20 mg/ml DA, with only 45% of the larvae reaching pupation at 20 mg/ml DA (Table 7). In addition, although not completely akinetic, there is a dramatic reduction in larval movement on media supplemented with 20 mg/ml DA. On lower DA concentrations, Catsup mutant phenotypes associated with late larval and pupal ELPs are induced (see Figure 4). Development of larvae is slowed at concentrations of 3 mg/ml DA or greater, with a delay in pupation of ~24–36 hr at 3 mg/ml DA. Larvae that do undergo pupation are undersized, with the pupal cases tanning much more rapidly and significantly darker relative to the controls. In addition, most pupae that die start autolysis but cannot undergo metamorphosis, producing empty or partially empty pupal cases. Some unhatched pupae exhibit melanotic patches, although this effect is not severe. However, not all phenotypes were induced by exogenously supplemented DA. DA-supplemented, wild-type larvae did not exhibit melanotic salivary glands or pseudotumors at any concentration tested.

We also tested the effects of DA by transferring adults to media supplemented with 1, 2, 5, and 10 mg/ml of DA. The DA supplementation did not have any effect on adult viability and fertility. However, as observed in the first series of experiments, progeny raised in DA-supplemented media resulted in similar effects on larval viability, developmental rates, and late larval and pupal phenotypes. In fact, the effects of DA supplementation were enhanced in this second series of experiments. For example, only 66% of larval progeny from the second series of experiments pupated when raised in media containing 2 mg/ml DA, as compared to 86% of larvae pupating from the first series of experiments. In addition, a low incidence of melanotic tumors were observed in larval progeny of DA-fed adults, with the proportions exhibiting tumors increasing with increasing DA concentrations. The tumors varied in size from very tiny to very large, ultimately filling as much as one half of the pharate adult bodies in the pupal cases. Those with very large tumors did not eclose.

To examine the effects of DA on fertility, adult progeny reared on media supplemented with DA were transferred to unsupplemented media and tested for decreased fecundity and progeny viability. Eggs laid per female during the first 48 hr after eclosion were reduced from 33 to 67% relative to controls. However, these differences largely disappeared by the third day. Thus, effects on egg deposition were seen in adults newly removed from DA, but these deficiencies were not permanent. Furthermore, embryo viability and hatching rates of these eggs were identical to those of control females.

We conclude from these studies that most of the phenotypes observed in Catsup mutants can be directly attributed to increases in catecholamine levels, as each phenotype was observed to some degree in wild-type individuals reared in DA-supplemented medium.

Catsup mutants exhibit elevated tyrosine hydroxylase activity: The depletion of free tyrosine and the concomitant increase in catecholamine levels in Catsup embryos suggest that mutations may result in the hyperactivation of TH. To examine this possibility, we measured TH activity in Catsup mutants. We chose for this analysis Catsup¹¹ (GI) and Catsup¹² (GIV) mutants, both of which were subjected to the analysis of catecholamine pools described above, and Catsup¹ (GV) and Catsup²⁶ (GV), presumptive null mutations. Assays were performed at the late third larval instar white prepupal stage. Here, TH activity is at sufficiently high levels in wild-type strains to allow for reliable and consistent measurements.

These assays demonstrate that Catsup mutations are indeed associated with significant increases in TH activity (Table 8). Because it is a weak allele, ~75% of Catsup¹¹ homozygous flies survive, allowing us to assay TH activity in these mutants. Mean TH activity in homozygous Catsup¹¹ mutant larvae is fourfold higher than wild-type levels. Consistent with other dominant characteristics of Catsup mutations, Catsup¹¹ balanced heterozygotes also demonstrate significantly higher TH activity, showing more than a fivefold increase relative to controls. Surprisingly, TH activity does not display a dose-response relative to the dose of the Catsup¹¹ allele, indicating that the effect of Catsup¹¹ on TH activity is a fully dominant trait. TH activity in Catsup¹, Catsup¹², and Catsup²⁶ balanced heterozygotes is also significantly elevated. Catsup¹/CyO larvae exhibit an approximately fourfold increase and Catsup¹²/CyO larvae more than a sevenfold increase in TH activity. To verify the effects of the presumptive null allele Catsup¹ on TH activity, we assayed TH activity in a newly generated null mutation, Catsup²⁶. Balanced heterozygotes of Catsup²⁶ exhibit a fourfold increase in TH activity (Table 8), a level similar to the increases detected in Catsup¹ mutants. These data demonstrate that both weak and severe mutations of Catsup dramatically increase TH activity, and they suggest that GV lethal alleles behave genetically and biochemically as null mutations.

TH-hyperactivating Catsup alleles are loss-of-function mutations: Genetic complementation, phenotypic analysis, and TH assays indicate that Catsup alleles are loss-of-function mutations and that GV lethal alleles are null mutations. To verify these findings, we conducted Western blot analysis with anti-Catsup antibody to determine whether Catsup mutants show reduced protein levels. The anti-Catsup antibody detects only a single band of ~50 kD, a size consistent with the predicted molecular mass of Catsup (STATHAKIS 1998 ).

Third instar larvae protein extracts of Catsup¹¹ (GI), Catsup¹ (GV), Catsup¹² (GIV), and Catsup²⁶ (GV) balanced heterozygotes all exhibit a clear reduction in the level of Catsup protein (Figure 8). This finding is consistent with our genetic analysis that indicates all three of these alleles are either severe or null mutations (Table 1). Surprisingly, the amount of protein detected by anti-Catsup antiserum in Catsup¹ (GV), Catsup¹² (GIV), and Catsup²⁶ (GV) mutants is ~16–22% of the level relative to wild-type controls (Figure 8). Because the Catsup mutant larvae assayed were all balanced heterozygotes, protein nulls would be expected to express 50% of the wild-type Catsup protein. The apparent >50% reduction in protein levels suggests that the dominant characteristics of these mutations might be explained by the destabilization of wild-type Catsup protein in these mutants or possibly the perturbation of an autoregulating mechanism normally involved in Catsup function. On the other hand, protein extracts from mutants for the weak allele Catsup¹¹ (GI) do not exhibit a visible reduction in Catsup protein levels. Interestingly, the band of immunoreactivity observed in the Catsup¹¹ lane is broader relative to all other lanes and may reflect the presence of a second smaller band (Figure 8). This observation is consistent with sequence analysis of Catsup¹¹ mutants, which identified a frameshift mutation that prematurely truncates the Catsup protein at its carboxyl end (STATHAKIS 1998 ). This analysis indicates that at least some Catsup mutations, including two severe GV lethals, are associated with a reduction in the amount of Catsup gene product and confirms that these alleles are loss-of-function mutations. In turn, this analysis suggests that the hyperactivation of TH appears to be caused by a reduction in the function of the Catsup gene product.

View larger version (49K): [in this window] [in a new window] [Download PPT slide]   Figure 8. Western blot analysis of Catsup and TH expression in Catsup mutants. Autoradiographs of a Western blot probed sequentially with an anti-Catsup antiserum, an anti-TH antibody, and an anti--tubulin antibody. Except where noted, each lane contains 5.0 µg of total protein extracted from third instar larvae of the following strains: Oregon-R (lane 1), 1/2 Oregon-R, 2.5 µg total protein loaded (lane 2), Catsup1/CyO, a GV lethal (lane 3), Catsup11/CyO, a GI lethal (lane 4), Catsup12/CyO, a GIV lethal (lane 5), and Catsup26/CyO, a GV lethal (lane 6). Top, anti-Catsup guinea pig polyclonal antiserum detects a single band of ~50 kD with relative band ODs as follows: Oregon-R, 1.00; 1/2 Oregon-R, 0.90; Catsup1/CyO, 0.22; Catsup11/CyO, 1.83; Catsup12/CyO, 0.20; Catsup26/CyO, 0.16. This band is significantly reduced in Catsup1, Catsup12, and Catsup26 mutants relative to the Oregon-R controls. Middle, an anti-TH rabbit polyclonal antibody detects a single band of ~58 kD with relative band ODs as follows: Oregon-R, 1.00; 1/2 Oregon-R, 0.33; Catsup1/CyO, 0.74; Catsup11/CyO, 0.61; Catsup12/CyO, 1.39; Catsup26/CyO, 0.35. The amount of TH protein detected in each of the Catsup mutants is equivalent to or less than that observed in Oregon-R larvae. Bottom, an anti--tubulin antibody detects a band of ~35 kD. This protein serves as a control for protein loads and to establish that general degradation of the proteins was not occurring.

Hyperactivation of TH enzyme activity is not associated with an increase in TH protein: TH activity is controlled by a complex array of mechanisms that act at the level of gene expression, mRNA translation, and protein modification. As a first step in assessing how Catsup might regulate TH activity, we determined whether TH protein levels were altered in Catsup mutants. Western blots of late third instar larval proteins were probed with an anti-TH antibody. The expected 58-kD TH protein is detected in the mutant strains Catsup¹, Catsup¹¹, Catsup¹², and Catsup²⁶ (Figure 8). In none of these was the TH protein level elevated; hence, the increase in TH activity in Catsup mutants cannot be attributed to increased production or stabilization of the TH protein. In fact, for Catsup¹ (GV), Catsup¹¹ (GI), and Catsup²⁶ (GV) mutants, the opposite effect is observed because the amount of TH protein detected is only 35–74% of the level found in wild-type controls (Figure 8). In the course of our experiments, we found that the activity of TH in Catsup mutants is much less stable than in wild-type strains. It would appear that the less stable activity is accompanied by a loss of the protein itself.

To verify these findings, we repeated the Western blot experiments but varied the amount of protein extract loaded by one-half to twofold. While immunoreactive signal intensity increased or decreased as we increased or decreased the amount of protein extract loaded, we never detected increased levels of TH protein in Catsup mutants. In all cases, TH protein seems to be reduced below wild-type levels. Similar results were obtained when we repeated the Western blot experiments but varied the concentration of anti-TH mouse monoclonal primary antibody (1:250 to 1:2500) or anti-mouse IgG-HRP-conjugated secondary antiserum (1:1000 to 1:5000). These data indicate that our detection method demonstrates a linear response under the conditions used. Taken together, this analysis indicates that Catsup mutations do not affect TH protein levels and suggests that the elevated TH activity observed in Catsup mutants is most likely attributed to a post-translational mechanism that alters TH catalytic activity.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Catsup mutations hyperactivate TH activity by a posttranslational mechanism:
Our data demonstrate that Catsup mutations result in a significant elevation in TH activity that corresponds to increased catecholamine pool levels. Catsup¹ (GV), Catsup¹¹ (GI), Catsup¹² (GIV), and Catsup²⁶ (GV) mutant third instar larvae show a four- to sevenfold increase in TH activity (Table 8). Because Catsup mutants exhibit three broad overlapping ELPs with associated cuticle defects or catecholamine-related abnormalities (Table 4), we infer that TH activity is hyperactivated throughout development. The significantly elevated catecholamine pool levels measured in Catsup mutant embryos (Table 6) and pupae (WRIGHT 1996 ) are consistent with elevated TH activity at these developmental time points.

Our data also suggest that the regulation of TH activity by the Catsup gene product is achieved through a posttranslational mechanism. Western blot analysis of protein extracts from Catsup mutant third instar larvae shows TH protein levels that are at or below wild-type levels, indicating that TH synthesis is not elevated in these mutants (Figure 8). A wide range of anti-TH antibody dilutions and protein concentrations were tested extensively to verify that our detection of TH protein was in a linear range. We are confident that a four- to sevenfold increase in the amount of TH protein corresponding to the elevated degree of TH activity would have been detected. We therefore conclude that the wild-type Catsup gene product acts posttranslationally to modify the catalytic activity of TH, and that mutations in this gene result in significant hyperactivation of TH activity.

Catsup functions as a negative regulator of tyrosine hydroxylase activity:
Two alternative genetic mechanisms could explain the hyperactivation of TH activity and catecholamine biosynthesis in Catsup mutants. Catsup alleles might be gain-of-function mutations in a gene that is normally required to activate TH activity. Such mutations could either be neomorphic, causing inappropriately timed or mislocalized TH activity, or hypermorphic mutations resulting in the hyperactivation of TH. Alternatively, these Catsup alleles could represent loss-of-function or hypomorphic mutations in a gene that normally functions as a negative regulator of TH activity. Several lines of evidence support the conclusion that the Catsup alleles described in this report are loss-of-function mutations, and that the most severe class, the GV lethals, are null mutations. First, with the exception of GV lethal alleles, Catsup hemizygotes produce stronger phenotypes than homozygotes, as would be expected for hypomorphic alleles (Table 1). Second, the penetrance of weaker alleles is increased to similar levels when in trans to either GV lethal alleles of Catsup or to de